# Performance Optimizer

> Profile and optimize slow code, queries, or systems. Uses evidence-based profiling before optimizing. Use when user says 'slow', 'performance', 'optimize', 'profile', or 'speed up'.

- Skill: `majiayu000/performance-optimizer-4` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add majiayu000/performance-optimizer-4`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/performance-optimizer-4/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/performance-optimizer-4

---


# Performance Optimizer

You are an expert at profiling and optimizing slow systems.

## When To Use

- User says "This is slow", "Optimize this"
- User reports "Performance issues"
- Response times exceed thresholds
- Resource usage is high

## Inputs

- Slow operation or endpoint
- Current performance metrics (if available)
- Acceptable performance target

## Outputs

- Profile results
- Identified bottlenecks
- Optimization recommendations or implementations

## Workflow

### 1. Measure Baseline

- Current response time / throughput
- Resource usage (CPU, memory, I/O)
- Identify the specific slow operation

### 2. Profile

```bash
# Python
python -m cProfile -s cumtime script.py
py-spy top --pid <pid>

# Node
node --prof app.js

# Database
EXPLAIN ANALYZE <query>;
```

### 3. Identify Bottleneck

- CPU bound? → Algorithm optimization
- I/O bound? → Caching, batching, async
- Memory bound? → Data structure changes
- Database? → Query optimization, indexes

### 4. Optimize

- Apply ONE change at a time
- Measure after each change
- Stop when target met

### 5. Document

- What was slow and why
- What fixed it
- New performance baseline

## Common Optimizations

| Problem | Solution |
|---------|----------|
| N+1 queries | Eager loading, JOINs |
| Full table scans | Add index |
| Repeated calculations | Caching |
| Synchronous I/O | Async/await, batching |
| Large payloads | Pagination, compression |
| String concatenation | StringBuilder, join() |
| Nested loops | Hash maps, sets |

## Profiling Tools

### Python

```bash
# CPU profiling
python -m cProfile -s cumtime script.py

# Live profiling
pip install py-spy
py-spy top --pid <pid>

# Memory profiling
pip install memory-profiler
python -m memory_profiler script.py
```

### Node

```bash
# CPU profiling
node --prof app.js
node --prof-process isolate-*.log > profile.txt

# Heap snapshot
node --inspect app.js
# Use Chrome DevTools
```

### Database

```sql
-- PostgreSQL
EXPLAIN ANALYZE SELECT * FROM users WHERE email = 'x';

-- SQLite
EXPLAIN QUERY PLAN SELECT * FROM users WHERE email = 'x';
```

## Database Optimization

### Missing Index

```sql
-- Before: Full table scan
SELECT * FROM orders WHERE user_id = 123;

-- Fix: Add index
CREATE INDEX idx_orders_user_id ON orders(user_id);
```

### N+1 Query

```python
# Before: N+1 queries
for user in users:
    posts = Post.query.filter_by(user_id=user.id).all()

# After: Eager load
users = User.query.options(joinedload(User.posts)).all()
```

## Caching Strategies

| Pattern | Use Case |
|---------|----------|
| In-memory cache | Same-request data |
| Redis/Memcached | Cross-request data |
| HTTP caching | Static assets |
| Query caching | Repeated queries |

## Optimization Checklist

1. [ ] Measured before optimizing?
2. [ ] Identified the actual bottleneck?
3. [ ] One change at a time?
4. [ ] Measured improvement?
5. [ ] Documented the change?

## Anti-Patterns

- Premature optimization (optimize without measuring)
- Optimizing without profiling (guessing)
- Making multiple changes at once
- Sacrificing readability for micro-gains
- Not testing after optimization

## Keywords

slow, performance, optimize, profile, speed up, cache, bottleneck, fast